Entertainment
Anthropic tests AI’s capacity for sabotage
As the hype around generative AI continues to build, the need for robust safety regulations is only becoming more clear.
Now Anthropic—the company behind Claude AI—is looking at how its models could deceive or sabotage users. Anthropic just dropped a paper laying out their approach.
Sam Altman steps down as head of OpenAI’s safety group
Anthropic’s latest research — titled “Sabotage Evaluations for Frontier Models” — comes from its Alignment Science team, driven by the company’s “Responsible Scaling” policy.
The goal is to gauge just how capable AI might be at misleading users or even “subverting the systems we put in place to oversee them.” The study focuses on four specific tactics: Human Decision Sabotage, Code Sabotage, Sandbagging, and Undermining Oversight.
Think of users who push ChatGPT to the limit, trying to coax it into generating inappropriate content or graphic images. These tests are all about ensuring that the AI can’t be tricked into breaking its own rules.
Mashable Light Speed
In the paper, Anthropic says its objective is to be ready for the possibility that AI could evolve into something with dangerous capabilities. So they put their Claude 3 Opus and 3.5 Sonnet models through a series of tests, designed to evaluate and enhance their safety protocols.
The Human Decision test focused on examining how AI could potentially manipulate human decision-making. The second test, Code Sabotage, analyzed whether AI could subtly introduce bugs into coding databases. Stronger AI models actually led to stronger defenses against these kinds of vulnerabilities.
The remaining tests — Sandbagging and Undermining Oversight — explored whether the AI could conceal its true capabilities or bypass safety mechanisms embedded within the system.
For now, Anthropic’s research concludes that current AI models pose a low risk, at least in terms of these malicious capabilities.
“Minimal mitigations are currently sufficient to address sabotage risks,” the team writes, but “more realistic evaluations and stronger mitigations seem likely to be necessary soon as capabilities improve.”
Translation: watch out, world.
Topics
Artificial Intelligence
Cybersecurity
-
Entertainment4 weeks ago
CES 2025: How to buy (and save $390 on) the Dreame X50 Ultra robot vacuum
-
Entertainment2 weeks ago
China’s DeepSeek AI might be smarter than OpenAI’s smartest AI
-
Entertainment4 weeks ago
8 mind-blowing gadgets at CES 2025 I need in my life
-
News1 week ago
DeepSeek AI gets hit with data privacy red flag by Italy and Ireland
-
Games4 weeks ago
How To Run Sega CD Games On Steam Deck
-
Entertainment4 weeks ago
How to live stream the Dallas Mavericks in 2025
-
Entertainment4 weeks ago
ChatGPT isn’t responsible for the Los Angeles fires, but it does use a crazy amount of water
-
Entertainment4 weeks ago
Best AI laptops at CES 2025, from Asus to Lenovo